|
1
|
Wang K, Song L, Wang C, Ren Z, Zhao G, Dou J, Di J, Barbastathis G, Zhou R, Zhao J, Lam EY. On the use of deep learning for phase recovery. LIGHT, SCIENCE & APPLICATIONS 2024; 13:4. [PMID: 38161203 PMCID: PMC10758000 DOI: 10.1038/s41377-023-01340-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Received: 07/31/2023] [Revised: 11/13/2023] [Accepted: 11/16/2023] [Indexed: 01/03/2024]
Abstract
Phase recovery (PR) refers to calculating the phase of the light field from its intensity measurements. As exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, PR is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. In recent years, deep learning (DL), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various PR problems. In this review, we first briefly introduce conventional methods for PR. Then, we review how DL provides support for PR from the following three stages, namely, pre-processing, in-processing, and post-processing. We also review how DL is used in phase image processing. Finally, we summarize the work in DL for PR and provide an outlook on how to better use DL to improve the reliability and efficiency of PR. Furthermore, we present a live-updating resource ( https://github.com/kqwang/phase-recovery ) for readers to learn more about PR.
Collapse
Affiliation(s)
- Kaiqiang Wang
- Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China.
- School of Physical Science and Technology, Northwestern Polytechnical University, Xi'an, China.
- Department of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China.
| | - Li Song
- Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China
| | - Chutian Wang
- Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China
| | - Zhenbo Ren
- School of Physical Science and Technology, Northwestern Polytechnical University, Xi'an, China
| | - Guangyuan Zhao
- Department of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China
| | - Jiazhen Dou
- School of Information Engineering, Guangdong University of Technology, Guangzhou, China
| | - Jianglei Di
- School of Information Engineering, Guangdong University of Technology, Guangzhou, China
| | - George Barbastathis
- Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA
| | - Renjie Zhou
- Department of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China
| | - Jianlin Zhao
- School of Physical Science and Technology, Northwestern Polytechnical University, Xi'an, China.
| | - Edmund Y Lam
- Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China.
| |
Collapse
|
|
2
|
Sha J, Qiu W, He G, Luo Z, Huang B. Improving the resolution of Fourier ptychographic imaging using an a priori neural network. OPTICS LETTERS 2023; 48:6316-6319. [PMID: 38039256 DOI: 10.1364/ol.508134] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Received: 10/12/2023] [Accepted: 11/15/2023] [Indexed: 12/03/2023]
Abstract
In this paper, we propose a dual-structured prior neural network model that independently restores both the amplitude and phase image using a random latent code for Fourier ptychography (FP). We demonstrate that the inherent prior information within the neural network can generate super-resolution images with a resolution that exceeds the combined numerical aperture of the FP system. This method circumvents the need for a large labeled dataset. The training process is guided by an appropriate forward physical model. We validate the effectiveness of our approach through simulations and experimental data. The results suggest that integrating image prior information with system-collected data is a potentially effective approach for improving the resolution of FP systems.
Collapse
|